Driver-assistance systems are used in order to support the driver or user of a vehicle in different traffic situations. For example, driver-assistance systems can warn the driver of looming collisions or perform automatic braking if a dangerous traffic situation has been detected. These collision warning systems essentially consider vehicles or objects that are either standing on the road or are moving on the road in the same direction as the own vehicle. The known driver-assistance systems are also already able to take oncoming traffic into account. For example, this is the case in what is known as the left-turn assistant, which implements automatic emergency braking if an acute collision danger with an oncoming vehicle is determined in a left turn. A passing assistant, on the other hand, warns the driver if a collision danger with oncoming traffic exits during a passing maneuver. This system, too, ultimately carries out automatic braking if the driver does not abandon the risky passing maneuver.
A basic problem inherent in the warning driver-assistance functions is that said warnings are to be output only if the driver is in actual need of support. The reason for this is that excessive warnings may annoy the driver and ultimately lead to a deactivation of the system by a user intervention. In the same way, a driver-assistance system that outputs such warnings too infrequently in situations in which the vehicle driver needs assistance, results in similar irritation of the vehicle driver.
Predicting the movement of the own and a second vehicle over multiple seconds typically constitutes a very significant challenge for a driver-assistance system. If the prediction is incorrect, the driving situation is used incorrectly and an incorrect system response is executed or fails to be executed as a result (false positive or false negative). The concepts for predicting the vehicle movements used heretofore are based on simple movement models, e.g., the vehicle is moving along a straight line or a circular path at a constant speed or acceleration, which may possibly allow no precise prediction of the traffic situation and therefore lead to incorrect system reactions with relative frequency. This may result in unintended system reactions, especially when bottlenecks and oncoming traffic are encountered.